PR #17544: Fix ConvNeXt classifier activation bug

Imported from GitHub PR #17544

Fixes the issue #17386.

Please find the bug, fix, and test [gist here](https://colab.research.google.com/gist/Frightera/93a784fc0e03a19f5804e7e14bfbecba/-17386.ipynb).
Copybara import of the project:

--
eedabb6 by Kaan <[email protected]>:

Pass classifier_activation arg to "Head"
--
e226091 by Kaan <[email protected]>:

Add classifier_activation to the docstring
--
3856fed by Kaan <[email protected]>:

Add classifier_activation unit test
--
c3dfc34 by Kaan <[email protected]>:

Move classifier_activation validation before head creation
--
71eaa69 by Kaan <[email protected]>:

Update test_application_classifier_activation
--
ebd6940 by Kaan Bıçakcı <[email protected]>:

Reformatting using format.sh

--
bce4ac9 by Kaan Bıçakcı <[email protected]>:

Fix test_application_classifier_activation test

--
00d4889 by Kaan Bıçakcı <[email protected]>:

Fix Head params to accept classifier_activation

--
2c22d37 by Kaan Bıçakcı <[email protected]>:

Revert "Fix Head params to accept classifier_activation"

This reverts commit 00d4889.

--
3abd441 by Kaan Bıçakcı <[email protected]>:

Exclude RegNet in test_application_classifier_activation

Merging this change closes #17544

FUTURE_COPYBARA_INTEGRATE_REVIEW=#17544 from Frightera:frightera_fix_17386 3abd441
PiperOrigin-RevId: 509550999

keras/applications/applications_test.py

@@ -191,6 +191,16 @@ def test_application_pooling(self, app, last_dim):
         )
         self.assertShapeEqual(output_shape, (None, last_dim))
 
+    @parameterized.parameters(MODEL_LIST)
+    def test_application_classifier_activation(self, app, _):
+        if "RegNet" in app.__name__:
+            self.skipTest("RegNet models do not support classifier activation")
+        model = app(
+            weights=None, include_top=True, classifier_activation="softmax"
+        )
+        last_layer_act = model.layers[-1].activation.__name__
+        self.assertEqual(last_layer_act, "softmax")
+
     @parameterized.parameters(*MODEL_LIST_NO_NASNET)
     def test_application_variable_input_channels(self, app, last_dim):
         if backend.image_data_format() == "channels_first":

keras/applications/convnext.py

@@ -324,11 +324,12 @@ def apply(x):
     return apply
 
 
-def Head(num_classes=1000, name=None):
+def Head(num_classes=1000, classifier_activation=None, name=None):
     """Implementation of classification head of RegNet.
 
     Args:
       num_classes: number of classes for Dense layer
+      classifier_activation: activation function for the Dense layer
       name: name prefix
 
     Returns:
@@ -342,7 +343,11 @@ def apply(x):
         x = layers.LayerNormalization(
             epsilon=1e-6, name=name + "_head_layernorm"
         )(x)
-        x = layers.Dense(num_classes, name=name + "_head_dense")(x)
+        x = layers.Dense(
+            num_classes,
+            activation=classifier_activation,
+            name=name + "_head_dense",
+        )(x)
         return x
 
     return apply
@@ -522,8 +527,12 @@ def ConvNeXt(
         cur += depths[i]
 
     if include_top:
-        x = Head(num_classes=classes, name=model_name)(x)
         imagenet_utils.validate_activation(classifier_activation, weights)
+        x = Head(
+            num_classes=classes,
+            classifier_activation=classifier_activation,
+            name=model_name,
+        )(x)
 
     else:
         if pooling == "avg":